Most chemotherapy for leukemia is limited by its toxicity to both leukemia cells and normal tissue. Ideally, therapy would target bio-molecules which are essential for leukemia stem-cell (LSC) but not normal hematopoietic stem-cell (HSC) survival. We demonstrate that the DNA methyl-transferase enzyme DNMT1 constitutes such an ideal molecular target; in HSC, DNMT1 is necessary for the self-renewal gene repression that must precede lineage-specific differentiation. In contrast, in LSC, DNMT1 is aberrantly recruited to repress pro-differentiation genes, prevent terminal differentiation and maintain dysregulated proliferation. Therefore, intermittent depletion of DNMT1 with non-DNA damaging doses of the cytosine analogue decitabine terminally differentiates LSC but increases self-renewal of HSC. This constitutes a very favorable therapeutic index. In the first aim of this proposal, we translate these observations into clinical practice. Decitabine was originally developed as a DNA-damaging agent. Doses were escalated to maximum tolerated levels in traditional phase I studies. Current regimens of decitabine still employ relatively high doses and drug administration is cycled to allow the patient to recover from toxicity. However, we show that effective DNMT1 depletion can be produced with levels of decitabine that do not damage DNA. Therefore, we propose lowering the dose of decitabine to minimize/avoid DNA damage and toxicity while maintaining DNMT1 depletion. The lack of toxicity will allow weekly, multi-year therapy to sustain the differential effect on LSC and HSC and introduces the possibility of adjuvant, maintenance or cancer prevention applications. We have demonstrated the remarkable clinical effectiveness and tolerability of using decitabine in this way in the treatment of severe sickle cell disease, where decitabine is administered 1-3X/week for multi-year treatment durations 2,3 4. In Specific Aim 1 of this proposal, we seek proof of concept of this regimen in treating malignancy. We believe the effort described in Aim 1 will be a significant advance in the treatment of malignancy. However, obstacles to realizing the full clinical potential of DNMT1 depletion by decitabine remain: (i) pharmacogenomic variation in cytidine deaminase (CDA), the enzyme which breaks-down decitabine, produces significant inter-individual variation in pharmacokinetics (PK) and pharmacodynamics (PD), compromising the ability to predict clinical effects in response to a specific dose; (ii) because of CDA- mediated drug destruction, decitabine has limited oral bioavailability, a significant impediment to the proposed treatment paradigm of multi-year, chronic therapy; (iii) we show that the major tumor stratagem for resistance to decitabine is CDA-mediated destruction of the drug. Such resistance may be especially likely in the intended chronic low-dose application of decitabine. In Specific Aim 2, we propose to surmount all three obstacles by combining decitabine with the CDA inhibitor tetrahydrouridine (THU) in a single oral formulation.